Early history
Over more than 4 decades, CUCC has developed methods for handling such information. Refinements in data management were made necessary by improved quantity and quality of survey; but refinements in data management also helped to drive those improvements. The first CUCC Austria expedition, in 1976, produced only Grade 1 survey for the most part (ref Cambridge Underground 1977 report).
In the 1980s, the use of programmable calculators to calculate survey point position from compass, tape, and clinometer values helped convince expedition members to conduct precise surveys of every cave encountered. Previously, such work required hours of slide rule or log table work. On several expeditions, such processing was completed after the expedition by a FORTRAN program running on shared mainframe time. BASIC programs running on personal computers took over with the release of the BBC Micro and then the Acorn A4. A full history of this period is described in Taking Expo Bullshit into the 21st Century - a story of the data management system up to Spring 1996. [This was less than five years after Tim Berners-Lee published the world's very first web page on 6th August 1991. So the expo website is nearly as old as the web itself.]
Survex - cave surveying
In the 1990s, Olly Betts and Wookey began work on "Survex", a program in C for the calculation and 3-D visualization of centerlines, with intelligent loop closure processing. Julian Todd's Java program "Tunnel" facilitated the production of attractive, computerized passage sketches from Survex centerline data and scanned hand-drawn notes. A history of survex article covering the period 1988-1996 was published in Cambridge Underground 1996.
Before Survex, in 1990 we used Sean Kelly's Surveyor '88, written for the Queen Mary College Belize Expedition, and in the 1980s we used Andy Waddington's SU 5.13C, a fortran programme on the IBM mainframe at the university (and, probably, at the UK Atomic Energy Establishment at Windscale). The routines to produce graphical output on a pen-plotter were ported to the University Computing Service by Philip Sargent in 1984. (Over Christmas 1983, the university mainframe had its RAM doubled: from 16 to 32 MB.)
1990 was the first year we had a computer on expo. Before that we had only programmable calculators. Unfortunately there were 3-inch and 5-inch floppy disc mismatches and it was an Acorn Archimedes anyway [Which had an ARM chip, not an Intel one. Yes younglings, we had ARM chips in those days too.]. See the 1990 report.
Initial cave data management
Along with centrelines and sketches, descriptions of caves were also affected by improvements in data management. In a crucial breakthrough, Andrew Waddinton introduced the use of the nascent markup language HTML to create an interlinked, navigable system of descriptions (see "Expo Bullshit"). Links in HTML documents could mimic the branched and often circular structure of the caves themselves. For example, the reader could now follow a link out of the main passage into a side passage, and then be linked back into the main passage description at the point where the side passage rejoined the main passage. This elegant use of technology enabled and encouraged expedition members to better document their exploration.
To organize all other data, such as lists of caves and their explorers, expedition members eventually wrote a number of scripts which took spreadsheets (or comma separated value files, .CSV ) of information and produced webpages in HTML. Other scripts also used information from Survex data files. Web pages for each cave as well as the indexes which listed all of the caves were generated by one particularly powerful script, make-indxal4.pl . The same data was used to generate a prospecting map as a JPEG image. The system of automatically generating webpages from data files reduced the need for repetitive manual HTML coding. Centralized storage of all caves in a large .CSV file with a cave on each row made the storage of new information more straightforward.
HTML and the website
From having a set of HTML files, it was a small step to publish a website. The CUCC Expo Website, which publishes the cave data, was originally created by
Andy Waddington in the early 1990s and was hosted by Wookey.
1999 scripts and spreadsheets
The version control system was CVS. The whole site was just static HTML, carefully
designed to be RISCOS-compatible (hence the short 10-character filenames)
as both Wadders and Wookey were RISCOS" people then (in the early 1990s).
Wadders wrote a huge amount of info collecting expo history, photos, cave data etc.
Martin Green added the survtab.csv file to contain tabulated data for many caves around 1999, and a script to generate the index pages from it. Dave Loeffler added scripts and programs to generate the prospecting maps in 2004. The server moved to Mark Shinwell's machine in the early 2000s, and the version control system was updated to subversion.
Breaking out of spreadsheet cells into HTML
Not all types of data could be stored in spreadsheets or survey files. In order a display descriptions on the webpage for an individual cave, the entire description, written in HTML, had to be typed into a spreadsheet cell. A spreadsheet cell makes for an extremely awkward HTML editing environment. To work around this project, descriptions for large caves were written manually as a tree of HTML pages and then the main cave page only contained a link to them.
A less obvious but more deeply rooted problem was the lack of relational information. One table named folk.csv stored names of all expedition members, the years in which they were present, and a link to a biography page. This was great for displaying a table of members by expedition year, but what if you wanted to display a list of people who wrote in the logbook about a certain cave in a certain expedition year? Theoretically, all of the necessary information to produce that list has been recorded in the logbook, but there is no way to access it because there is no connection between the person's name in folk.csv and the entries he wrote in the logbook.
The only way that relational information was stored in our csv files was by putting references to other files into spreadsheet cells. For example, there was a column in the main cave spreadsheet, cavetab2.csv , which contained the path to the QM list for each cave. The haphazard nature of the development of the "script and spreadsheet" method meant that every cave had an individual system for storing QMs. Without a standard system, it was sometimes unclear how to correctly enter data.
From " Troggle: a revised system for cave data management", by Philip Sargent and Aaron Curtis, CUCC [with some additions]. Original (2006) paper: " Troggle: a novel system for cave exploration information management", by Aaron Curtis.
HTML and browser incompatibilities
This was a much bigger problem in the past than it is possible to imagine now (most of the early editing is done on an Acorn platform using !Zap). So much so that we have a separate page all about it: "Platform portability - making the website work widely".
Version control
Another important element of this system was version control. The entire data structure was stored initially in a Concurrent Version System repository, and later migrated to Subversion [now using git in 2020]. Any edits to the spreadsheets which caused the scripts to fail, breaking the website, could be easily reversed.
2006 to 2011 - and troggle
In 2006 Aaron Curtis decided that a more modern set of generated, database-based pages made sense, and so wrote Troggle. This uses Django to generate pages. This reads in all the logbooks and surveys and provides a nice way to access them, and enter new data.From the 2009 expo report:
-
This year's expedition also had a non-caving goal (not just drinking Gösser). Recently [since 2006] members of CUCC have started to develop a piece of software called Troggle, which aims to facilitate keeping track of logbook entries, typing up surveys, caves etc, and save time in a lot of the work that goes on behind the scenes when expo is over. This year was the first time Troggle would be tested "in the field" (well, spud hut).
After Expo 2009 the version control system was updated to a DVCS (Mercurial, aka 'hg'),
-
'because a distributed version control system makes a great deal of sense for expo
(where it goes offline for a month or two and nearly all the year's edits happen)'.
The site was moved to Julian Todd's 'seagrass' server (in 2010), but the change from a 32-bit to 64-bit machine broke the website autogeneration code, which was only fixed in early 2011, allowing the move to complete.
By 2011 Troggle was under development with a wiki hosted on the CUCC server and we have a snapshot of the status in April 2011. As you can see, Troggle was still very incomplete in 2011.
The handbook was separate for a while until Martin Green added code to merge the old static pages and new troggle dynamic pages into the same site. This is now the live system running everything (in 2022). Work on developing Troggle further still continues (see Troggle intro).
The data was split into separate repositories: the website, troggle, the survey data, the tunnel data. 'Seagrass' was turned off at the end of 2013, and the site moved to a machine managed by Sam Wenham at the university computer service in Feb 2014.
Some text taken from " Troggle: a revised system for cave data management", by Philip Sargent and Aaron Curtis, CUCC [with some additions]. Original (2006) paper: " Troggle: a novel system for cave exploration information management", by Aaron Curtis.
2018 Four repositories
In 2018 we had 4 repositories, 2 mercurial, 2 git
- loser - the survex cave survey data (hg)
- drawings - the tunnel and therion cave data and drawings (git)
- expoweb - the website pages, handbook, generation scripts (hg)
- troggle - the database/software part of the survey data management system - see notes on troggle for further explanation (git)
In spring 2018 Sam, Wookey and Paul Fox updated the Linux version and the Django version (i.e. troggle) to something vaguely acceptable to the university computing service and fixed all the problems that were then observed.
Early 2019 - Leaving the Computing service
In early 2019 the university computing service upgraded its firewall rules which took the server offline completely.
Wookey eventually managed to find us free space (a virtual machine) on a debian mirror server somewhere in Leicestershire (we think). This move to a different secure server means that all ssh to the server now needs to use cryptographic keys tied to individual machines. There is an expo-nerds email list (all mailing lists are now hosted on wookware.org as the university list system restricted what non-Raven-users could do) to coordinate server fettling.
At the beginning of the 2019 expo two repos had been moved from mercurial to git: troggle and drawings (formerly called tunneldata).
Wookey: July 2019
The troggle software has been migrated to git, and the old erebus and cvs branches (pre 2010) removed. Some decrufting was done to get rid of log files, old copies of embedded javascript (codemirror, jquery etc) and some fat images no longer used.
The tunneldata repo has also been migrated to git, and renamed 'drawings' as it includes therion data too these days.
The loser repo and expoweb repo need more care in hg->git migration (expoweb is the website content - which is published by troggle). Loser should have the old 1999-2004 CVS history restored, and maybe Tom's annual snapshots from before that, so ancient history can usefully be researched (sometimes useful). It's also a good idea to add the 2015, 2016 and 2017 ARGE data we got (in 2017) added in the correct years so that it's possible to go back to an 'end of this year' checkout and get an accurate view of what was found (for making plots and length stats). All of that requires some history rewriting, which is best done at the time of conversion.
Similarly expoweb is full of bloat from fat images and surveys and one 82MB thesis that got checked in and then removed. Clearing that out is a good idea. I have a set of 'unused fat blob' lists which can be stripped out with git-gilter. It's not hard to make a 'do the conversion' script, ready for sometime after expo 2019 has calmed down.
May 2020 and django
Wookey has now moved 'expoweb' from mercurial to git largely "as-is". Mark Shinwell has said that he will help on the loser (survex files) migration to git.
In May we were on django 1.7 and python 2.7.17. Sam continued to work on upgrading django from v1.7 . We wanted to upgrade django as quickly as possible because old versions of django had unpatched security issues. [Upgrading to later django versions is a real pig - not helped by the fact that all the tools to help do it are now out of date for these very old django releases.]
- "Django 1.11 is the last version to support Python 2.7. Support for Django 1.11 ends in 2020." see: django versions. You will notice that we are really outstaying our welcome here, especially as python2.7 was declared dead in January this year.
- For a table displaying the various versions of django and support expiry dates see the django download page. Django 1.7 expired in December 2015. Django: full deprecation timeline.
- Ubuntu 20.04 came out on 23rd April but it does not support python2 at all. So we cannot use it for software maintenance (well be can, but only using non-recommended software, which is what we are trying to get away from).
We planned to upgrade from django 1.7 to django 1.11, then port from python2 to python3 on the same version of django, then upgrade to as recent a version of django as we could. But we have discovered that django1.7 works just fine with python3, so we will move the development version to python3 during June and then upgrade the server operating system from Debian stretch to buster before tackling the next step: thinking deeply about when we migrate from django to something else.
Enforced time at home under covid lockdown is giving us a new impetus to writing and restructuring the documentation for everything.
June 2020
Sam was a bit overworked in trying to get an entire university to work remotely during Covid lockdown so Philip [Sargent] started on the python2/3 conversion and got troggle on django 1.7 to work on python 3.5 and then 3.8. He then did the slog of migrating it through the django versions up to 1.11.29 - the last version before django 2.0 . 1.11.29 is an LTS (long term support) version of django. In doing this we had to retreat to python3.7 due to a django plugin incompatibility.
In the course of these migrations several unused or partly-used django plugins were dropped as they caused migration problems (notably staticfiles) and the plug-ins pillow, django-registration, six and sqlparse were brought up to recent versions. This was all done with pip in a python venv (virtual environment) on a Windows 10 machine running ubuntu 20.04 under WSL (Windows Systems for Linux) v1.
Missing troggle functions were repaired and partly-implemented pages, such as the list of all cavers and their surveyed passages, were finished and made to work. The logbook parsing acquired a cacheing system to re-load pre-parsed files. The survex file parsing was completely rebuilt to reduce the excessive memory footprint. While doing so the parser was extended to cover nearly the full range of survex syntax and modified to parse, but not store, all the survey stations locations. A great many unused classes and some partly written code ideas were deleted.
July 2020
Wookey upgraded debian on the server from 9 stretch to 10 buster and we got the python3 development of troggle running as the public version (with some http:// and https:// glitches) by 23rd July. Buster will be in-support definitely until June 2024 so we are rather pleased to be on a "not ancient" version of the operating system at last. This concided with a last tweak at improving the full cave data file import so now it runs on the development system in ~80 seconds. Which is considerably more useful than the ~5 hours it was taking earlier this year.
April 2021
Covid lockdown has been good to troggle. During March and April Philip migrated troggle up to Django 2.2.19, excising the ancient and unused user registration system on the way. Django 2.2 LTS is a long-term stable relase which will be in-support by Django until April next year. Wookey discovered and ran the Django system testsuite on the Debian server thus enabling us to use a necessary (but obstensibly outdated) link between Django and the database MariaDB. As of April 9th troggle is now running on software which is actually 'in date'.
We plan to stick with this configuration for a year.
March 2022
On 15th March Wookey upgraded the server to the debian release 11 bullseye. At this point 'debian stable' is bullseye and has python 3.9 as standard. We will quickly migrate to Django 3.2 LTS which is now a year old and which will be supported until April 2024. Bullseye will be in support until June 2026.
Django migrations are not nearly as painful as they used to be, and troggle is already compatible with Django 4.0.3 (though we won't use that). So the presure to migrate from Django is now very greatly lessened. However, see troggle architecture speculations and possible migration from Django.
We should not need to anything until we move from Django 3.2 LTS to 4.2 LTS before April 2024.
July 2022
Wookey at last, after much effort, got the loser repository converted from mercurial to git, with much tidying and history-reconfabulation. He says this will need to be done again, but it was good enough for the 2022 Expo. Also the troggle code was changed: survex files edited on a webpage now automatically commit to git with no user involvement.
Just before expo, we finished integrating the formerly-separate 'wallets' script. So now the progress of scanning and tunneling survey data can be managed more easily. This has turned out to be unexpectedly powerful. And it works on the data back to 1999 too.
Wookey's replacment WiFi antenna (previously we had used Sam's) turned out to be useless, so internet access was not available in the hut this year. This was a real pain.
More recent
For the current situation see expo systems status.